Title :
Adaptive learning control system based on the designed multilevel stochastic supervisor of an artificial neural network
Author_Institution :
Dept. of Chem. Eng., Helsinki Univ., Finland
Abstract :
The aim of this work is to present the design of a learning control system composed of a hierarchical stochastic reinforcement scheme with penalty to supervise and improve the convergence of a artificial feedforward neural network based on a backpropagation algorithm. The performance of this ANN system is tested while applied to the command and optimization of two different chemical processes: a continuous stirred-tank reactor with a control objective to follow the given operating set points, and a pulsed liquid-liquid extraction column with a control objective to maintain the column within its optimal operating zone
Keywords :
adaptive control; backpropagation; chemical industry; feedforward neural nets; intelligent control; neurocontrollers; process control; stochastic systems; adaptive learning control; backpropagation; chemical processes; continuous stirred-tank reactor; convergence; feedforward neural network; multilevel stochastic supervisor; pulsed liquid-liquid extraction column; stochastic systems; Adaptive control; Adaptive systems; Algorithm design and analysis; Artificial neural networks; Control systems; Convergence; Neural networks; Optimal control; Programmable control; Stochastic systems;
Conference_Titel :
Neural Networks, 1995. Proceedings., IEEE International Conference on
Conference_Location :
Perth, WA
Print_ISBN :
0-7803-2768-3
DOI :
10.1109/ICNN.1995.487756
